DANDI:000945

Using two-way ANOVA, we found that anesthesia is a significant factor in the delayed response for both RSU ... and FSU

Results 3.2, paragraph 2; Discussion, paragraph beginning 'Although these results demonstrate'

What the paper reports

• RSU: two-way ANOVA, anesthesia factor, p < 0.001
• FSU: two-way ANOVA, anesthesia factor, p < 0.01

With one recording per rat and condition, either without the `_2` recordings (Rep2) or with the earliest of each pair (Rep3), anesthesia is a factor for RSU at p = 4.2e-09 to 1.8e-08 and for FSU at p = 0.0012 to 0.0015 under A1/A2 and W1/W2, and the anesthetized mean is lower in both (RSU 99.7 against 104.3 to 104.6%, FSU 99.7 against 101.2 to 101.3%). With every recording (Rep1) RSU holds (p = 3.6e-06 to 4.3e-06) but FSU gives p = 0.038 to 0.041, which misses the 0.01 bound; that reading gives partially reproduced. Leaving out the recording filed under both BH498 and BH508 gives FSU p = 0.0016 to 0.0019 under Rep2 and Rep3, and 0.048 to 0.052 under Rep1 (above 0.05 under W2), where the RSU part alone still holds and the criterion gives partially reproduced ("when one cell type holds"). With every recording the claim is thus partial at best, so the reproduced outcome rests on the assumption. The anesthetized units come from two rats, so the state effect may also be a rat effect. The question is recorded for the paper's authors.

How the test differed from the paper’s method

• Assumed that the paper's analysis used one recording per rat and condition. The paper mentions no repeats, Methods 2.3 fits one recording per parameter set per rat without stating it ("Recordings for a single rat for each set of parameters were completed on the same day" says when a rat's recordings were made, so it supports the assumption only weakly), and the dandiset's description says "Some subjects have duplicate recordings taken a few months after the original ones". Which recording of a pair is the original is open, as three `_2` recordings are earlier than their pair; both choices (Rep2, Rep3) give the same outcome.
• Anesthetized units from two rats.

Delayed responses from FSUs showed a significantly higher response to 60% DC over 6% ... with 6% DC found to be not significantly different from sham

Results 3.3, paragraph 1; Figure 7D

What the paper reports

• Kruskal-Wallis with post hoc Wilcoxon, p < 0.001, Bonferroni-corrected

Parts (1) and (2) hold under every reading, without a 30% group included, so they do not depend on the substitute; the Kruskal-Wallis is significant (p = 0.0051 without a 30% group, 0.011 C1, 4.5e-07 to 1.4e-06 C2) and 60% is above 6% (p = 0.0051 alone, corrected p = 0.015 over three pairs). Part (3), 6% against sham, cannot be tested, so the claim is partially reproduced, the outcome the criterion gives when (1) and (2) hold without (3). The difference is small and both groups are below baseline (99.6% and 98.2%).

How the test differed from the paper’s method

• No sham recording; part (3) is reported as not tested.
• The 30% group is a substitute (C1 or C2) or absent (C0); it enters only the Kruskal-Wallis.

In the anesthetized model, these delayed responses had an average onset latency of 0.24 s ± 0.08 s, while in the awake model they had an average onset latency of 0.51 s ± 0.14 s ... Comparing the distributions using the Wilcoxon rank-sum test, we found that delayed response onset times varied significantly between awake and anesthetized models

Results 3.1, paragraph 1; Figure 4H and 4I

What the paper reports

• Wilcoxon rank-sum test, p < 0.001 (n = FSU with a significant delayed response at 1500 Hz PRF, awake and anesthetized (counts not given))
• as printed in the Figure 4 legend: Wilcoxon rank sum test, p < 10⁻⁴

Under none of the 24 readings (S1/S2, D1/D2, L1/L2, Rep1/Rep2/Rep3) are awake onsets later at p < 0.001; rank-sum p is 0.018 to 0.92, and in 8 readings the anesthetized mean is the later. Awake means are 0.40 to 0.58 s (within 0.51 +- 0.14), anesthetized means 0.38 to 0.52 s (all outside 0.24 +- 0.08). Measured values are the S1 D1 L1 Rep1 means (anesthetized, awake). 193 to 265 of 306 awake FSU (86% under S1 D1) and 34 or 39 of 50 anesthetized FSU have a delayed response under the paper's rule as the claims file reads it, and latencies spread over 0.07 to 1.39 s in both states, so the rule flags most units and the onsets look largely noise-driven. The anesthetized FSU come from two rats (BH498, BH512). Q1 and Q2 select the same recordings here.

How the test differed from the paper’s method

• The anesthetized FSU come from two labelled rats where the paper had ten; batch 2 is not in the dandiset.
• T2 (505 trials) cannot be run; the files hold 500 trials, or fewer in five recordings.

Examining the time-locked response from RSU in the awake model using Kruskal–Wallis ANOVA, we find that PRF is a significant factor ... Post hoc tests show that 30 Hz and 300 Hz PRF induced responses lower than all higher PRFs ... The sham condition was significantly different than all but the 30 Hz condition ... 1500 Hz, 3000 Hz, and 4500 Hz showed no significant differences

Results 3.2, paragraph 1; Figure 5A

What the paper reports

• PRF as a factor: Kruskal-Wallis, p < 0.001
• 30 and 300 Hz against each higher PRF: post hoc Wilcoxon, p < 0.001, Bonferroni-corrected
• sham against 300, 1500, 3000 and 4500 Hz: post hoc Wilcoxon, p < 0.001, Bonferroni-corrected

PRF is a significant factor (Kruskal-Wallis p = 3.3e-16, W1 Rep1) but the post hoc pattern differs. The highest mean is at 3000 Hz (108.2%; 1500 Hz 98.5%), and 3000 Hz differs from 1500 and 4500 Hz, so part (4) fails under every reading; part (3) fails under every reading; part (2) holds only under W2 Rep1 and W2 Rep3. Parts (5) and (6) need the sham and were not tested; even if both held, part (1) with at most three others falls short of the four the partial outcome needs.

How the test differed from the paper’s method

• No sham recording exists; parts (5) and (6) are reported as not tested, and a claim with an untested part is not judged reproduced.
• Ten awake recordings repeat a condition in the same rat; results are given with them (Rep1), without the `_2` recordings (Rep2) and with the earliest recording of each pair (Rep3).

The anesthetized model had similar statistical results, with the only difference being that no significant difference was observed between 30 Hz and 300 Hz PRF levels

Results 3.2, paragraph 1; Figure 5C

The Kruskal-Wallis is significant (p = 7.3e-08 W1, 7.7e-07 W2) and 30 and 300 Hz do not differ (part 7), but parts (2), (3) and (4) fail under both readings. 1500 Hz is the peak (107.0%) and differs from 4500 Hz; 30 and 300 Hz are not below 3000 or 4500 Hz. Even if the two sham parts held, part (1) with three others falls short of the five the partial outcome needs. With two rats PRF and rat are confounded, as the 30 Hz group is BH498 alone and the 1500 Hz peak comes from BH512 (110.4% against 101.7% in BH498, whose means span 100.0 to 101.7%).

How the test differed from the paper’s method

• Two labelled rats (BH498 at five PRFs, BH512 without 30 Hz) where the paper had ten; BH521's anesthetized recordings have no labels.
• Parts (5) and (6) need the sham and were not tested.

Two-way ANOVA additionally finds that anesthesia is a significant factor in the time-locked response from RSU ... Anesthesia was observed to decrease the average magnitude of RSU responses

Results 3.2, paragraph 1; Discussion, paragraph beginning 'Although these results demonstrate'

What the paper reports

• two-way ANOVA, anesthesia factor, p < 0.01

Anesthesia is a factor at p = 0.0083 to 0.018 over the twelve readings (A1/A2, W1/W2, Rep1/Rep2/Rep3), below 0.01 only under W1 Rep3, and in the reverse direction under every reading; the anesthetized mean (102.6 to 102.9%) is above the awake mean (100.1 to 100.4%). The anesthetized units come from two rats and the awake units from seven, so the state effect is partly a rat effect; with only BH498 and BH512 it disappears (p = 0.425).

How the test differed from the paper’s method

• Anesthetized units from two rats; awake units from seven.

FSU in both the awake and the anesthetized model show no significant differences due to PRF levels

Results 3.2, paragraph 1; Figure 5B and 5D

By the paper's test FSU time-locked responses differ across PRF in both states, Kruskal-Wallis p = 1.8e-12 awake and 5.5e-04 anesthetized (W1 Rep1), with 4 to 6 significant post hoc pairs awake and 2 or 3 anesthetized under every reading. The effects are small; group medians span 98.7 to 103.4% awake and 99.8 to 103.9% anesthetized, and the awake 3000 Hz mean is 106.8%. G2 (with sham) cannot be run. The anesthetized groups come from two rats.

FSUs in the awake model show the strongest responses to 4500 Hz PRF, with significant differences between 30 Hz and the highest three PRFs, and 300 Hz and the highest two PRFs ... 30 Hz and 300 Hz were not found to be significantly higher than sham

Results 3.2, paragraph 2; Figure 5F

What the paper reports

• post hoc Wilcoxon, p < 0.001, Bonferroni-corrected

PRF is a significant factor (p = 4.6e-12 Rep1) and 30 Hz differs from 1500, 3000 and 4500 Hz (part 3, tested without direction as the criterion says "differs"), but the strongest mean is at 30 Hz (103.5%, 4500 Hz 100.3 to 100.4%), the opposite of the paper's pattern, and 300 Hz does not differ from 4500 Hz (corrected p = 1.0), under every repeat reading. Part (5) needs the sham and was not tested; even if it held, part (1) with two others falls short of the three the partial outcome needs.

PRF was found to be a significant factor for RSU and FSU in both awake and anesthetized settings ... Both RSUs and FSUs in the anesthetized rats had similar sensitivity as their time-locked responses, broadly preferring higher PRFs

Results 3.2, paragraph 2; Figure 5G and 5H

PRF is not a significant factor for anesthetized delayed responses, Kruskal-Wallis p = 0.16 (W1) and 0.12 (W2) for RSU, 0.16 and 0.18 for FSU. The higher PRFs have the higher pooled mean by under one point (RSU 99.97 against 99.15%, FSU 99.87 against 99.35%). Every rat and PRF mean is within 98.5 to 100.9%, and the 30 Hz group is BH498 alone.

How the test differed from the paper’s method

• Two labelled rats where the paper had ten.

FSU's show no time-locked response, and their delayed response increases almost linearly with I SPTA .

Results 3.2, paragraph 4; Figure 6F and 6H

The time-locked part fails (fig5bd, awake Kruskal-Wallis p = 1.8e-12). Delayed FSU means are 103.5, 98.7, 101.8, 100.6 and 100.3% from 30 to 4500 Hz (Spearman rho = -0.4, Pearson r = -0.27, Rep1; rho = -0.3 Rep2 and Rep3).

Although it is not statistically significant, we found that 1500 Hz produces the highest mean time-locked response

Results 3.2, paragraph 4; Figure 5A

The highest mean is at 3000 Hz under every reading (107.7 to 108.7%), significantly above 1500 Hz (98.1 to 98.7%; corrected p = 0.0040 or less).

Based on mean values, 60% DC appears to produce the strongest time-locked response, although it is only statistically significantly higher than the 6% condition.

Results 3.3, paragraph 1; Figure 7A

The highest mean is at 6% under every reading (176.3 to 180.8%; 60% 147.1 to 147.5%), which does not need the 30% group. With the substitute 30% groups, 60% is significantly above 30% (corrected p < 1e-16).

How the test differed from the paper’s method

• The 30% group is a substitute (C1 or C2) or absent (C0), as for fig7a.

no time-locked response was observed as in the PRF study

Results 3.3, paragraph 1; Figure 7B

FSU time-locked responses differ across DC under every reading, and the 6% against 60% difference does not need the 30% group (Kruskal-Wallis p = 0.0054 without it; p = 3.5e-12 or smaller with a substitute, all three pairs significant). FSU respond strongly at 6% and 60% (means 156.8 and 205.3%, medians 115.0 and 128.6%) and not at the substitute 30% (median 100.8%), most in BH506 (338.6 and 403.3%).

How the test differed from the paper’s method

• The 30% group is a substitute (C1 or C2) or absent (C0), as for fig7a.

K -means clustering of detected neurons from one animal's recordings, showing 28 fast spiking units (FSU) and 32 regular spiking units (RSU), for 60 total.

Figure 3B and 3C; Methods 2.5

Both the reproduced and the partially reproduced outcome of the criterion require the clusters to separate on the initial phase and after-hyperpolarization durations, and the files hold no waveforms, so K1 and K2 cannot run and neither outcome can be reached. With the dataset's labels (K0) two single recordings (R1) give all three counts within tolerance, BH527_30_200 with 28 FSU, 34 RSU and 62 units (given as measured) and BH508_4500_200_2 with 28 FSU, 30 RSU and 58 units; under Rep2 and Rep3 only BH527_30_200 does. No rat's series matches (R2; totals 77 to 616). The match is no more than chance, as 22 of 59 recordings have a total within 60 +- 6 and a binomial model with the pooled FSU share (0.505) expects 3.84 recordings to match all three counts, against 2 observed. FSU have the higher median baseline rate in 46 of 59 labelled recordings (8.60 against 5.84 Hz over units).

How the test differed from the paper’s method

• Cell types are the dataset's labels (K0); that they are the paper's k-means output is assumed, and is consistent with the baseline rates without being shown.
• The other claims depend on this one for cell types. They use the K0 labels, as the claims file's header provides when the dataset carries labels, and do not need the waveform features whose absence blocks this claim, so they are tested.

We observed time-locked responses from RSU in both awake and anesthetized recordings, with an observable increase in firing rate during tFUS stimulation

Results 3.1, paragraph 1; Figure 4C and 4E

The criterion compares the 1500 Hz, 200 us RSU time-locked responses with the sham sessions of each state, and the dandiset holds no sham recording. The awake half already fails the condition that the 1500 Hz mean be above 100%, as RSU at 1500 Hz average 98.1 to 98.7% of baseline awake under every reading (n = 227 or 303), so the claim could not be reproduced whatever a sham would show; at best it would be partially reproduced, on the anesthetized half (106.4 to 107.0%, n = 78, two rats).

The awake model had a wider variance of onset latencies, and appears to have two peaks

Results 3.1, paragraph 1; Figure 4H

The claims file's readings give different results, and its header blocks a claim whose readings disagree. The awake SD exceeds the anesthetized SD under every D1 reading (0.384 to 0.396 s against 0.367 to 0.372 s) but is smaller under 10 of 12 D2 readings (0.381 to 0.393 s against 0.393 to 0.406 s) and larger by under 0.001 s in the other two. Under every reading the SDs are within 0.03 s of each other, with no contrast like the paper's 0.14 against 0.08 s. The awake histogram (S1 D1 L1 Rep1) has its largest count in the first allowed bin (61 of 262 at 0.071 s) and several smaller local maxima. Whether the first bin allowed is one starting at or after 67 ms (D1) or one ending after it (D2) is asked of the paper's authors.

RSUs in the awake rats had highest responses to 1500 Hz and 3000 Hz ... with a decrease at both higher and lower frequencies, and all conditions found to be significantly higher than sham

Results 3.2, paragraph 2; Figure 5E

What the paper reports

• 1500 and 3000 Hz highest: post hoc Wilcoxon, p < 0.001, Bonferroni-corrected
• each PRF against sham: post hoc Wilcoxon, p < 0.001, Bonferroni-corrected

PRF is a significant factor (p = 6e-20 Rep1, 4.5e-15 Rep2, 5.1e-15 Rep3) but the two highest means are 1500 and 30 Hz (106.7 and 106.6% Rep1), and 3000 Hz (102.3 to 103.3%) is below 30 Hz (corrected p < 1e-4); part (2) fails under both of its readings and every repeat reading. Part (3), every PRF above sham, needs the sham and cannot be tested. The criterion gives partially reproduced when (1) and (3) hold without (2), so the outcome would be partial if a sham held and not reproduced otherwise, and the data present do not decide between the two. The claim is therefore blocked as data missing, as fig4-rsu-time-locked-both-states is.

We observe that RSU's in both their time-locked and delayed responses initially increase till around 60 mW cm −2 , after which the response declines.

Results 3.2, paragraph 4; Figure 6E and 6G

The delayed means do not rise to their peak under any reading (106.6% at 30 Hz, 101.2 to 102.8% at 300 Hz, 106.7 to 108.8% at 1500 Hz), so at best the claim is partially reproduced, on the time-locked window. Time-locked RSU means rise from 30 Hz (96.4 to 96.8%) to a peak at 3000 Hz (60.26 mW/cm2, 107.7 to 108.7%) and fall at 4500 Hz (98.0 to 99.5%) under W2 with every repeat reading and under W1 with Rep1 and Rep3, but under W1 Rep2 the 300 Hz mean (98.2%) is 0.1 point above the 1500 Hz mean (98.1%, SEMs 1.1 and 1.3). With one recording per rat and condition the claims file's readings W1 and W2 thus disagree under Rep2, and its header blocks a claim whose readings disagree. How a window off the bin edges was measured is asked of the paper's authors.

Time-locked responses from RSUs showed a significantly lower response to 6% DC than those to 60%, with all conditions significantly higher than sham

Results 3.3, paragraph 1; Figure 7A and its legend

What the paper reports

• Kruskal-Wallis with post hoc Wilcoxon, p < 0.001, Bonferroni-corrected

Part (2) does not need the 30% group and fails under every reading, as 6% is the highest group (180.8%; 60% 147.5%) and does not differ from 60% (corrected p = 1.0; 0.67 to 0.91 without a 30% group). Part (1), the Kruskal-Wallis over 6, 30 and 60%, needs the 30% group (p = 0.67 and 0.91 without it, 1.5e-22 or smaller with a substitute, so the substitutes alone decide it), part (4), 6% below 30%, needs it too, and part (3) needs the sham, so parts (1), (3) and (4) are untested. The criterion gives partially reproduced when (1) and at least one of (2) to (4) hold, so partial is reachable if (1) and (3) or (4) held, and the data present do not decide it. The claim is therefore blocked as data missing, as fig7c-rsu-delayed-dc is. The substitute 30% groups are far below both (90.4 to 98.7%). The 6% and 60% responses are largest in BH506 (249.0 and 209.9%); BH521 and BH526 give 113 to 121%.

How the test differed from the paper’s method

• The dandiset has no 30% DC recording of its own. The substitutes (C1, the PRF series' 1500 Hz, 200 us recordings of BH506 and BH526; C2, those of every rat, with each repeat reading) and C0, no 30% group, are shown; a part that needs the 30% group is treated as data missing.
• The 6% and 60% recordings come from three rats where the paper had ten.

Delayed responses however preferred 30% DC over 6% and 60% ... and 60% DC showed no significant difference from sham

Results 3.3, paragraph 1; Figure 7C and its legend

What the paper reports

• Kruskal-Wallis with post hoc Wilcoxon, p < 0.001, Bonferroni-corrected

Part (2), 30% above 6% and 60%, needs the paper's 30% DC recordings, which are not in the dandiset, and the substitutes decide it. With the same rats' 1500 Hz, 200 us recordings (C1) the Kruskal-Wallis is not significant (p = 0.21) and no pair differs; with every such recording (C2) the effect is significant (p = 2.2e-04 Rep1, 3.7e-05 Rep2, 1.9e-04 Rep3) and 30% is above 6% and 60%. Part (4), 6% against 60%, also depends on the 30% group through the Bonferroni family; 6% is below 60% at p = 0.024 alone and at corrected p = 0.071 over three pairs. Part (3) needs the sham. The C1 substitutes (BH506, BH526) were made three to five months before the same rats' DC recordings; C2 adds recordings from August 2023 to 6 March 2024, among them BH498's of 5 and 6 March, three to four days before the DC recordings of 9 March.

How the test differed from the paper’s method

• The 30% group is a substitute (C1 or C2) or absent (C0), as for fig7a.